Anthropic says it identified and blocked instances in which scientists used its Claude AI models for research the company believes could support biological weapons development. In its account, Anthropic cites uses related to sensitive virology, including “gain-of-function” type work and experiments involving pathogens that could have pandemic potential.
CBS News reports that Anthropic describes this as misuse of Claude “in ways that could support biological weapons development.” Engadget adds that the company provides extensive case studies detailing how current AI models have been used in problematic ways, framing the examples as evidence of potential misuse pathways.
The outlets differ mainly in emphasis: the New York Post highlights specific categories of work Anthropic flagged, CBS focuses on the company’s statement that it blocked researchers, and Engadget focuses on the breadth of Anthropic’s documented case studies. Across coverage, all describe Anthropic’s response—identifying misuse and blocking further access—along with its concern that certain research applications could be dual-use in the context of biological weapons.
